ASR Graph of Thoughts (GoT) Model Context Protocol (MCP) Server
The Advanced Scientific Research (ASR) Graph of Thoughts (GoT) MCP server is a highly efficient implementation of the Model Context Protocol (MCP) that allows for sophisticated reasoning workflows using graph-based representations.
Project Overview
This project implements a Model Context Protocol (MCP) server architecture that leverages a Graph of Thoughts approach to enhance AI reasoning capabilities. It can be connected to AI models or applications like Claude desktop app or API-based integrations.
Project Structure
asr-got-mcp/
├── docker-compose.yml # Docker Compose configuration for multi-container setup
├── Dockerfile # Docker configuration for the backend
├── requirements.txt # Python dependencies
├── src/ # Source code
│ ├── server.py # Main server implementation
│ ├── asr_got/ # Core ASR-GoT implementation
│ │ ├── core.py # Core functionality
│ │ ├── stages/ # Processing stages
│ │ │ ├── stage_1_initialization.py
│ │ │ ├── stage_2_decomposition.py
│ │ │ ├── stage_3_hypothesis.py
│ │ │ ├── stage_4_evidence.py
│ │ │ ├── stage_5_pruning.py
│ │ │ ├── stage_6_subgraph.py
│ │ │ ├── stage_7_composition.py
│ │ │ └── stage_8_reflection.py
│ │ ├── utils/ # Utility functions
│ │ └── models/ # Data models
│ └── api/ # API implementation
│ ├── routes.py # API routes
│ └── schema.py # API schemas
├── config/ # Configuration files
└── tests/ # Test suite
Running the Project with Docker
This project provides a multi-container Docker setup for both the Python backend (FastAPI) and the static JavaScript client. The setup uses Docker Compose for orchestration.
Project-Specific Docker Requirements
- Python Version: 3.13-slim (as specified in the backend Dockerfile)
- System Dependencies:
build-essential
,curl
(installed in the backend image) - Non-root Users: Both backend and client containers run as non-root users for security
- Virtual Environment: Python dependencies are installed in a virtual environment (
/app/.venv
) - Static Client: Served via nginx (alpine) in a separate container
Environment Variables
The backend service sets the following environment variables (see Dockerfile):
PYTHONUNBUFFERED=1
MCP_SERVER_PORT=8082
(the FastAPI server port)LOG_LEVEL=INFO
Note: If you need to override or add environment variables, you can uncomment and use the
env_file
option indocker-compose.yml
.
Exposed Ports
- Backend (python-app):
- Host:
8082
→ Container:8082
(FastAPI server)
- Host:
- Client (js-client):
- Host:
80
→ Container:80
(nginx static server)
- Host:
Build and Run Instructions
Build and start all services:
docker compose up --build
This will build both the backend and client images and start the containers.
Access the services:
- Backend API: http://localhost:8082
- Static Client: http://localhost/
Integration with AI Models
This MCP server can be integrated with:
- Claude desktop application
- API-based integrations with AI models
- Other MCP-compatible clients
Development
To set up a development environment without Docker:
- Clone this repository
- Create a virtual environment:
python -m venv venv
- Activate the virtual environment:
- Windows:
venvScriptsactivate
- Linux/Mac:
source venv/bin/activate
- Windows:
- Install dependencies:
pip install -r requirements.txt
- Run the server:
python src/server.py
License
This project is licensed under the Apache License 2.0 - see the LICENSE file for details.
If you update dependencies, remember to rebuild the images with docker compose build
.
ASR Graph of Thoughts Server
Project Details
- SaptaDey/Graph-of-Thought-MCP
- Apache License 2.0
- Last Updated: 5/7/2025
Recomended MCP Servers
Talk with Azure using MCP
An MCP implementation for Selenium WebDriver
An MCP server implementation that seamlessly connects Claude and other AI models to HubSpot CRM data. Provides tools...
free MCP server hosting using vercel
服务器管理仓库
A Model Context Protocol (MCP) server that sends webhook notifications when called.
This read-only MCP Server allows you to connect to Phoenix data from Claude Desktop through CData JDBC Drivers....
通过AI大模型操作安卓设备